Data Processing Improvements in MS Peak Detection for Trace Quantitation from Accurate Mass LC/MS Peptide Maps

Library Number:
Part Number:
BETH GILLECE-CASTRO; Marc V. Gorenstein; Daniel Golick; Keith Richardson; Barry Dyson; Scott Berger; Jeff Mazzeo; Thomas E. Wheat; Diane Diehl
Waters Corporation
Content Type:
Content Subtype:

Novel Aspect:

New algorithms have been developed for identification and quantitation of trace levels of protein variants.



The primary structure of a protein can be characterized by peptide mapping, and the same analytical technique can be used to identify modifications to the protein structure. Complete interpretation of complex LC/MS chromatograms with accurate mass measurement can be time-consuming and labor intensive.



Detection of m/z peaks in complex LC/MS data is theoretically straightforward. The issues of dynamic range, detector saturation, and appropriate matching can hinder accurate interpretation of peptide maps. A software tool that makes visual and quantitative comparisons can contribute to efficient data interpretation. We have developed prototype software to address the large data sets found in protein characterization laboratories. The

algorithm uses image analysis principles to identify signals in the large complex data sets found in peptide mapping. The peaks detected by the Apex3D algorithm and were then deconvoluted to collapse the multiplycharged isotope clusters to peptide molecular weights. Apex3D is specifically designed to interpret the resolved isotope clusters produced by oa-Tof mass spectrometers.


Preliminary Results:

Digests of proteins and glycoproteins were analyzed by LC/MS on oa-Tof mass spectrometers to test automated peak detection. Detected peaks were matched to known sequences and modifications including glycan structures. In a second experiment protein samples were subjected to forced degradation with peroxide or base to determine the oxidation state or deamidation state of the protein before and after treatment. Modified peptides were detected in the presence of native peptides. In a third experiment a protein was spiked with known amounts of variant peptides (0.2 – 10%) of peptide to test the limits of the software’s ability to detect trace components in a digest.


Title Format File Size
720002654en PDF 486.29kB